UNICEF Malnutrition Dataset Visualization

The dataset provides comprehensive information on malnutrition levels across countries and UNICEF regions. The data is sourced from the UNICEF website: https://data.unicef.org/resources/dataset/malnutrition-data/

# Import libraries
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

# Load the dataset
excel_url = 'path/malnutrition.xlsx'
malnutrition_df = pd.read_excel(excel_url)
malnutrition_df.head()

	ISO code	Country and areas	Survey year	Year*	United Nations Region	United Nations Sub-Region	SDG Region	UNICEF Region	UNICEF Sub-Region	WHO Region	...	Stunting	Stunting Footnote	WAZ Survey Sample (N)	Underweight	Underweight Footnote	Fieldwork period	Report Author	Source	Short Source	U5 Population ('000s)
0	AFG	AFGHANISTAN	1997	1997	Asia	Southern Asia	Central Asia and Southern Asia	SA	SA	EMRO	...	53.2	cw 20	4846.0	44.9	c	January,1997-May,1997	CIET International	Afghanistan 1997 multiple indicator baseline (...	MICS	3706.024902
1	AFG	AFGHANISTAN	2004	2004	Asia	Southern Asia	Central Asia and Southern Asia	SA	SA	EMRO	...	59.3	w 21	NaN	31.7	b	May,2004-June,2004	Ministry of Public Health (Afghanistan), UNICE...	Summary report of the national nutrition surve...	NNS	4705.370117
2	AFG	AFGHANISTAN	2013	2013	Asia	Southern Asia	Central Asia and Southern Asia	SA	SA	EMRO	...	40.4	r	4426469.0	24.6	s	May,2013-October,2013	Ministry of Public Health, UNICEF and the Aga ...	Afghanistan National Nutrition Survey 2013	NNS	5433.032227
3	AFG	AFGHANISTAN	2018	2018	Asia	Southern Asia	Central Asia and Southern Asia	SA	SA	EMRO	...	38.2	r	19539.2	19.1	s	March,2018-November,2018	KIT Royal Tropical Institute	Afghanistan Health Survey 2018	Other	6147.354980
4	ALB	ALBANIA	2005	2005	Europe	Southern Europe	Northern America and Europe	ECA	EECA	EURO	...	26.7	NaN	1090.2	6.6	NaN	October,2005-November,2005	Albanian National Institute of Statistics.	Albania multiple indicator cluster survey 2005...	MICS	214.872009

5 rows × 36 columns

#Find basic info of all columns
malnutrition_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1100 entries, 0 to 1099
Data columns (total 36 columns):
 #   Column                            Non-Null Count  Dtype  
---  ------                            --------------  -----  
 0   ISO code                          1100 non-null   object 
 1   Country and areas                 1100 non-null   object 
 2   Survey year                       1100 non-null   object 
 3   Year*                             1100 non-null   int64  
 4   United Nations Region             1100 non-null   object 
 5   United Nations Sub-Region         1100 non-null   object 
 6   SDG Region                        1100 non-null   object 
 7   UNICEF Region                     1088 non-null   object 
 8   UNICEF Sub-Region                 1088 non-null   object 
 9   WHO Region                        1091 non-null   object 
 10  World Bank Income Classification  1079 non-null   object 
 11  World Bank Region                 1100 non-null   object 
 12  LDC                               404 non-null    object 
 13  LIFD                              436 non-null    object 
 14  LLDC or SIDS                      389 non-null    object 
 15  UNICEF Survey ID                  1100 non-null   int64  
 16  WHO Global Database Number        1100 non-null   int64  
 17  Type of Estimate                  1100 non-null   object 
 18  WHZ Survey Sample (N)             982 non-null    float64
 19  Severe Wasting                    897 non-null    float64
 20  Severe Wasting Footnote           396 non-null    object 
 21  Wasting                           1055 non-null   float64
 22  Wasting Footnote                  549 non-null    object 
 23  Overweight                        964 non-null    float64
 24  Overweight Footnote               461 non-null    object 
 25  HAZ Survey Sample (N)             977 non-null    float64
 26  Stunting                          1051 non-null   float64
 27  Stunting Footnote                 466 non-null    object 
 28  WAZ Survey Sample (N)             891 non-null    float64
 29  Underweight                       1067 non-null   float64
 30  Underweight Footnote              517 non-null    object 
 31  Fieldwork period                  957 non-null    object 
 32  Report Author                     1098 non-null   object 
 33  Source                            1100 non-null   object 
 34  Short Source                      1100 non-null   object 
 35  U5 Population ('000s)             1100 non-null   float64
dtypes: float64(9), int64(3), object(24)
memory usage: 309.5+ KB

# Find null/NA columns
malnutrition_df.isnull().sum()

ISO code                              0
Country and areas                     0
Survey year                           0
Year*                                 0
United Nations Region                 0
United Nations Sub-Region             0
SDG Region                            0
UNICEF Region                        12
UNICEF Sub-Region                    12
WHO Region                            9
World Bank Income Classification     21
World Bank Region                     0
LDC                                 696
LIFD                                664
LLDC or SIDS                        711
UNICEF Survey ID                      0
WHO Global Database Number            0
Type of Estimate                      0
WHZ Survey Sample (N)               118
Severe Wasting                      203
Severe Wasting Footnote             704
Wasting                              45
Wasting Footnote                    551
Overweight                          136
Overweight Footnote                 639
HAZ Survey Sample (N)               123
Stunting                             49
Stunting Footnote                   634
WAZ Survey Sample (N)               209
Underweight                          33
Underweight Footnote                583
Fieldwork period                    143
Report Author                         2
Source                                0
Short Source                          0
U5 Population ('000s)                 0
dtype: int64

Handling null values for numeric columns

In the dataset, numerous null or NAN values are present. The proposed solution outlined below involves replacing the null values in numeric columns with their respective means. The process begins by grouping the dataframe by country, calculating the mean within each group, and subsequently filling the null values.

For countries with no values across any survey year, the corresponding columns are filled with zeros. This strategy was chosen over utilizing dropna() to avoid the removal of crucial data. The dropna() method resulted in a substantial reduction in the sample size, which was deemed undesirable.

# List of numeric columns
numeric_columns = malnutrition_df.select_dtypes(include='number').columns
# Group by 'Country and areas' and fill NaN values with the mean of each group for numeric columns
malnutrition_df[numeric_columns] = malnutrition_df.groupby(['Country and areas'])[numeric_columns].transform(lambda group: group.fillna(group.mean()))
# Fill remaining NaN values with 0 for numeric columns only
malnutrition_df[numeric_columns] = malnutrition_df[numeric_columns].fillna(0)

# Explore basic statistics
malnutrition_df.describe()

	Year*	UNICEF Survey ID	WHO Global Database Number	WHZ Survey Sample (N)	Severe Wasting	Wasting	Overweight	HAZ Survey Sample (N)	Stunting	WAZ Survey Sample (N)	Underweight	U5 Population ('000s)
count	1100.000000	1100.000000	1100.000000	1.100000e+03	1100.000000	1100.000000	1100.000000	1.100000e+03	1100.000000	1.100000e+03	1100.000000	1100.000000
mean	2005.846364	1366.496364	3100.658182	2.691202e+09	1.859558	6.615872	6.018644	2.709543e+09	27.187440	5.530074e+09	14.956497	6052.535110
std	9.576807	1804.161697	1689.987859	7.336579e+10	1.657911	4.886439	4.380790	7.386786e+10	15.877532	9.685860e+10	12.281906	16093.963919
min	1983.000000	1.000000	23.000000	0.000000e+00	0.000000	0.000000	0.000000	0.000000e+00	0.000000	0.000000e+00	0.000000	1.141000
25%	1998.000000	302.750000	2326.250000	2.914700e+03	0.700000	2.600000	2.600000	2.916000e+03	14.175000	3.349700e+03	4.300000	609.980255
50%	2007.000000	627.500000	3119.500000	5.929150e+03	1.400000	5.500000	5.200000	5.868750e+03	27.150000	6.768950e+03	12.450000	2023.768494
75%	2014.000000	2038.250000	3426.250000	1.886365e+04	2.575000	9.600000	8.300000	1.811731e+04	38.390625	2.383295e+04	21.900000	4265.401245
max	2022.000000	8994.000000	9862.000000	2.361127e+12	12.900000	25.300000	29.600000	2.377285e+12	73.600000	2.431224e+12	66.800000	131429.437500

Malnutriton factors

Stunting: Stunting is a form of malnutrition characterized by impaired growth and development, particularly in the early years of a child's life..

Wasting: Wasting refers to the rapid and severe loss of weight and muscle mass, often associated with acute malnutriti

Severe Wasting: Severe wasting is an even more critical condition than wasting. It indicates a severe degree of malnutrition, and children with severe wasting have a significantly low weight for their height.ons.

Underweight: Underweight is a general term indicating that an individual's weight is below the expected standard for theihealth.

Overweight: Overweight signifies an excess of body weight relative to height, often resulting from an imbalance between caloric intake and exl activity, and healthy lifestyle habits.

Analysis of Top 10 Countries with the Highest Average Malnutrition Factors Rate [Stunting, Overweight, Wasting, Underweight]

Factors Contributing to Elevated Malnutrition Rates:

1. Economic Challenges:

   • Underlying Issue: Widespread economic difficulties in these countries.
   • Impact on Malnutrition: Limited financial resources for families to afford nutritious food, leading to higher rates of stunting, wasting, and underweight.

2. Food Insecurity:

   • Underlying Issue: High prevalence of food insecurity.
   • Impact on Malnutrition: Persistent challenges in accessing a consistent, nutritious food supply contribute significantly to elevated malnutrition rates.

3. Healthcare Disparities:

   • Underlying Issue: Limited access to quality healthcare services.
   • Impact on Malnutrition: Hindered access to healthcare exacerbates health issues, heightening the risk of stunting, wasting, and underweight.

4. Prevalence of Infections:

   • Underlying Issue: Increased susceptibility to frequent infections.
   • Impact on Malnutrition: Frequent illnesses disrupt normal growth and development, playing a substantial role in the prevalence of stunting.

Understanding and addressing these underlying factors are crucial for developing targeted interventions aimed at reducing malnutrition rates in these regions. Comprehensive strategies encompassing economic development, food security measures, improved healthcare access, and infection prevention are essential to mitigate the impact of malnutrition on these communities.

1. Countries with the highest average stunting rates

top_countries = malnutrition_df.groupby("Country and areas")["Stunting"].mean().nlargest(10).index
y_values = malnutrition_df.groupby("Country and areas")["Stunting"].mean().round(2).loc[top_countries]

fig = px.bar(malnutrition_df, x=top_countries, y=y_values, color=top_countries, color_discrete_sequence=px.colors.sequential.Cividis)

fig.update_layout(
    title="Top 10 Countries with the highest average stunting rates",
    xaxis_title="Country name",
    yaxis_title="Average stunting rate",
    xaxis=dict(tickangle=-45)
)
fig.show()

2. Countries with the highest average overweight rates

top_countries_overweight = malnutrition_df.groupby('Country and areas')['Overweight'].mean().nlargest(10).index
y_values_overweight = malnutrition_df.groupby('Country and areas')['Overweight'].mean().round(2).loc[top_countries_overweight]

fig_overweight = px.bar(malnutrition_df, x=top_countries_overweight, y=y_values_overweight,color=top_countries_overweight, color_discrete_sequence=px.colors.sequential.Cividis)

fig_overweight.update_layout(
    title="Top 10 Countries with the highest average overweight rates",
    xaxis_title="Country name",
    yaxis_title="Average Overweight Rate",
    xaxis=dict(tickangle=-45),
)
fig_overweight.show()

3. Countries with the highest average wasting rates

top_countries_wasting = malnutrition_df.groupby('Country and areas')['Wasting'].mean().nlargest(10).index
y_values_wasting = malnutrition_df.groupby('Country and areas')['Wasting'].mean().round(2).loc[top_countries_wasting]

fig_wasting = px.bar(malnutrition_df, x=top_countries_wasting, y=y_values_wasting,color=top_countries_wasting, color_discrete_sequence=px.colors.sequential.Cividis)

fig_wasting.update_layout(
    title="Top 10 Countries with the highest average wasting rates",
    xaxis_title="Country name",
    yaxis_title="Average Wasting Rate",
    xaxis=dict(tickangle=-45),
)
fig_wasting.show()

4. Countries with the highest average underweight rates

top_countries_underweight = malnutrition_df.groupby('Country and areas')['Underweight'].mean().nlargest(10).index
y_values_underweight = malnutrition_df.groupby('Country and areas')['Underweight'].mean().round(2).loc[top_countries_underweight]

fig_underweight = px.bar(malnutrition_df,x=top_countries_underweight,y=y_values_underweight,color=top_countries_underweight, color_discrete_sequence=px.colors.sequential.Cividis)

fig_underweight.update_layout(
    title="Top 10 Countries with the highest average underweight rates",
    xaxis_title="Country name",
    yaxis_title="Average Underweight Rate",
    xaxis=dict(tickangle=-45),
)
fig_underweight.show()

5. Countries with the highest average malnutrition rates

• Eritrea: The average malnutrition rate in Eritrea is 26.56. This is likely due to a number of factors, including poverty, conflict, and drought.
• Timor-Leste: The average malnutrition rate in Timor-Leste is 28.76. This is also likely due to poverty, as well as a lack of access to clean water and sanitation.
• Bangladesh: The average malnutrition rate in Bangladesh is 27.2. This is due to a number of factors, including poverty, flooding, and a lack of access to education.
• India: The average malnutrition rate in India is 27.76. This is due to a number of factors, including poverty, a large population, and a lack of access to healthcare.
• Yemen: The average malnutrition rate in Yemen is 27.44. This is likely due to the ongoing civil war, which has caused widespread food insecurity.

Malnutrition is a serious problem that can have a number of negative consequences, including stunted growth, weakened immune systems, and increased vulnerability to disease. It is important to address the root causes of malnutrition in order to improve the lives of people in these countries.

# Calculate the average of the four columns
malnutrition_df['Overall malnutrition rate'] = malnutrition_df[['Overweight', 'Underweight', 'Wasting', 'Stunting']].mean(axis=1).round(2)

# Find the top 10 countries with the highest average
top_countries_average = malnutrition_df.groupby('Country and areas')['Overall malnutrition rate'].mean().nlargest(5).index
y_values_average = malnutrition_df.groupby('Country and areas')['Overall malnutrition rate'].mean().round(2).loc[top_countries_average]

fig_average = px.bar(
    malnutrition_df, 
    y=top_countries_average,  
    x=y_values_average,      
    color=top_countries_average,
    orientation='h',
    color_discrete_sequence=px.colors.sequential.Cividis
)

fig_average.update_layout(
    title="Top 5 Countries with the highest average malnutrition rates",
    yaxis_title="Country name",
    xaxis_title="Average Malnutrition Rate",
)

fig_average.show()

This map visualizes average malnutrition rates among children in different countries

# Create a choropleth map based on the average values
fig = px.choropleth(
    malnutrition_df,
    locations='ISO code',
    color='Overall malnutrition rate',
    hover_name='Country and areas',
    title='Average Malnutrition Indicator Across Countries',
    color_continuous_scale=px.colors.sequential.Cividis,
    labels={'Average': 'Average Malnutrition Indicator'}
)
fig.show()

Average Malnutrition Rate by World Bank Income Classification

• Low-income countries have the highest average malnutrition rate, at 38%. This is likely due to a number of factors, including poverty, conflict, and drought.
• Lower-middle-income countries have the second-highest average malnutrition rate, at 33%. This is also likely due to poverty, as well as a lack of access to clean water and sanitation.
• Upper-middle-income countries have an average malnutrition rate of 18.5%. This is lower than the two lower income groups, but still significantly higher than the high-income group.
• High-income countries have the lowest average malnutrition rate, at 10.4%. This is likely due to a number of factors, including wealth, access to food, and education.

As you can see, there is a strong correlation between income and malnutrition. The lower the income, the higher the malnutrition rate. This is a complex issue with no easy solutions, but it is important to be aware of the problem and to work towards solutions.

import plotly.express as px

# Calculate the average malnutrition rate for each income classification
avg_malnutrition_by_income = malnutrition_df.groupby('World Bank Income Classification')['Overall malnutrition rate'].mean().round(2).reset_index()

fig_pie_avg_by_income = px.pie(
    avg_malnutrition_by_income,
    names='World Bank Income Classification',
    values='Overall malnutrition rate',
    title='Average Malnutrition Rate by World Bank Income Classification',
    color_discrete_sequence=px.colors.sequential.Cividis
)
fig_pie_avg_by_income.show()

Malnutrition Trend Over Time

The following plot chart is a line graph that shows the trends in malnutrition rates over time for five different malnutrition categories: underweight, overweight, severe wasting, wasting, and stunting. The data is from 1983 to 2022.

• Overall, there has been a significant decline in malnutrition rates over time for all five categories. This is a positive trend that suggests that global efforts to reduce malnutrition are having an impact.
• The largest declines have been seen in underweight and severe wasting. These categories have both seen declines of more than 50% since 1985.
• Overweight rates have increased over time. This is a concerning trend, as overweight and obesity are also associated with a number of health problems.
• Wasting and stunting rates have also declined, but not as much as underweight and severe wasting. This suggests that there is still more work to be done in these areas.

malnutrition_cols = ['Severe Wasting', 'Wasting', 'Overweight', 'Stunting', 'Underweight']

fig = px.area(
    malnutrition_df,
    x='Year*',
    y=malnutrition_cols,
    title='Malnutrition Trends Over Time',
    labels={'value': 'Malnutrition Rates'},
    color_discrete_sequence=px.colors.sequential.Magma,
)

fig.update_layout(
    xaxis=dict(title='Year'),
    yaxis=dict(title='Malnutrition Rates'),
    xaxis_rangeslider_visible=True, 
).show()

Correlation Heatmap of Malnutritional Indicators

• Severe wasting and wasting are strongly positively correlated (0.83). This means that children who are severely wasted are also likely to be wasted. This is likely due to the fact that the same factors that contribute to severe wasting, such as poverty, lack of access to food, and repeated infections, also contribute to wasting.
• Wasting and underweight are also strongly positively correlated (0.83). This means that children who are wasted are also likely to be underweight. This is again likely due to the fact that the same factors that contribute to wasting also contribute to underweight.
• There is a weak negative correlation between overweight and stunting (-0.38). This means that children who are overweight are less likely to be stunted. This is a somewhat surprising finding, as overweight and stunting are often thought to be two sides of the same coin, known as the "double burden of malnutrition." However, it is possible that there is a protective effect of being overweight in early childhood, as it may provide a buffer against stunting later in life.
• There is a strong positive correlation between stunting and underweight (0.85). This means that children who are stunted are also likely to be underweight. This is likely due to the fact that the same factors that contribute to stunting, such as chronic malnutrition and repeated infections, also contribute to underweight.

Overall, the correlation matrix suggests that there is a strong relationship between different forms of malnutrition. This suggests that interventions to reduce malnutrition need to be comprehensive and address the underlying causes of multiple forms of malnutrition.

heatmap_data = malnutrition_df[['Severe Wasting', 'Wasting', 'Overweight', 'Stunting', 'Underweight']]

fig = px.imshow(heatmap_data.corr(), x=heatmap_data.columns, y=heatmap_data.columns,
                labels=dict(color='Correlation'), color_continuous_scale='Blues') 

fig.update_layout(title='Correlation Heatmap of Malnutritional Indicators')
fig.show()

Exploration of Malnutrition Indicator Relationships Across World Bank Regions

The following scatter matrix provides in-depth insights into the correlation between the given variables, offering a comprehensive view of malnutrition indicators across World Bank Regions.

fig = px.scatter_matrix(
    malnutrition_df,
    dimensions=['Severe Wasting', 'Wasting', 'Overweight', 'Stunting'],
    color='World Bank Region',
    title='Scatter Matrix of Malnutrition Indicators',
    color_discrete_sequence=px.colors.sequential.Plasma,
    height=1000, 
    width=1000    
)
fig.show()

Malnutrition Rates by UNICEF Region

• SA: South Asia
• ECA: Europe and Central Asia
• MENA: Middle East and North Africa
• SSA: Sub-Saharan Africa
• LAC: Latin America and the Caribbean
• EAP: East Asia and Pacific

The graph illustrates that Southern Asia has the highest rates of stunting, severe wasting, overweight, and wasting, with Sub-Saharan Africa following closely. In terms of overweight prevalence, Eastern and Central Africa (ECA) takes the lead, followed by the Middle East and North Africa (MENA).

fig = px.bar(
    malnutrition_df.groupby('UNICEF Region')[malnutrition_cols].mean().round(2).reset_index().melt(id_vars='UNICEF Region', var_name='Malnutrition Type', value_name='Malnutrition Rate'),
    x='UNICEF Region',
    y='Malnutrition Rate',
    color='Malnutrition Type',
    barmode='group',
    title='Malnutrition Rates by UNICEF Region',
    color_discrete_sequence=px.colors.sequential.Cividis
)
# Show the plot
fig.show()

Summary of Malnutrition factors for top 5 SDG regions with highest population of children under the age of 5

# Convert 'U5 Population (\'000s)' to millions
malnutrition_df['U5 Population (Millions)'] = malnutrition_df['U5 Population (\'000s)'] / 1000

# Group by SDG Region and calculate the sum of U5 Population in millions
total_sdg_population = malnutrition_df.groupby('SDG Region')['U5 Population (Millions)'].sum()

# Find the top 5 SDG Regions with the highest U5 Population
top_regions_df = malnutrition_df[malnutrition_df['SDG Region'].isin(total_sdg_population.nlargest(5).index)]

# Calculate the average malnutrition rate for each region
avg_malnutrition = top_regions_df.groupby('SDG Region')[malnutrition_cols + ['Overall malnutrition rate']].mean().reset_index()

# Create and show the table using Plotly
table = go.Figure(data=[go.Table(
    header=dict(values=['SDG Region'] + ['U5 Population (Millions)'] + malnutrition_cols + ['Overall malnutrition rate']),
    cells=dict(values=[avg_malnutrition['SDG Region']] 
               + [total_sdg_population.loc[avg_malnutrition['SDG Region']].round(2).tolist()] 
               + [avg_malnutrition[col].round(2).tolist() for col in malnutrition_cols 
               + ['Overall malnutrition rate']])
)]).show()

Comparison between malnutrition column pairs

1. Relationship between Wasting and Underweight Prevalence

Southern Asia exhibits the highest prevalence of both wasting and underweight. A consistent correlation is observed between wasting and underweight, indicating that nations with elevated wasting rates often experience concurrently high underweight rates. Nonetheless, certain exceptions exist within this pattern.

fig = px.scatter(
    malnutrition_df,
    x='Wasting',
    y='Underweight',
    color='United Nations Sub-Region',
    size='U5 Population (\'000s)',
    title='Correlation between Wasting and Underweight',
    labels={'Wasting': 'Wasting', 'Underweight': 'Underweight'},
    hover_data=['Country and areas', 'Survey year']
)
fig.show()

2. Correlation between Underweight and Stunting

Southern Asia exhibits the highest prevalence of both Underweight and Stunting. A consistent correlation is observed between Underweight and Stunting, indicating that nations with elevated Underweight rates often experience concurrently high Stunting rates. Nonetheless, certain exceptions exist within this pattern.

fig = px.scatter(
    malnutrition_df,
    x='Underweight',
    y='Stunting',
    color='United Nations Sub-Region',
    size='U5 Population (\'000s)',
    title='Correlation between Underweight and Stunting',
    labels={'Underweight': 'Underweight', 'Stunting': 'Stunting'},
    hover_data=['Country and areas', 'Survey year']
)
fig.show()

3. Correlation between Wasting and Stunting

Southern Asia exhibits the highest prevalence of both Wasting and Stunting. A consistent correlation is observed between Wasting and Stunting, indicating that nations with elevated Wasting rates often experience concurrently high Stunting rates. Nonetheless, certain exceptions exist within this pattern.

fig = px.scatter(
    malnutrition_df,
    x='Wasting',
    y='Stunting',
    color='United Nations Sub-Region',
    size='U5 Population (\'000s)',
    title='Correlation between Wasting and Stunting',
    labels={'Severe Wasting': 'Wasting', 'Stunting': 'Stunting'},
    hover_data=['Country and areas', 'Survey year']
)
fig.show()